as usual so fun and so much good intel and ideas all the time!! Love you guys! playing with chemosynthetic for sure and giving me such good monsters for halloween -haha!
Thank you Drew & Rory! Really appreciate you guys taking the time to dig into my question! Yes, you got it. I want to take an already created piece of clothing and switch it onto a MJ styled model in Lifestyle & Runway imagery. I tested this extensively. MJ won't do it, but other programs will. #juxtapositioning #juxtaposin'
Oh my gosh! What a start to your presentation. Sounds like you rocked it though! Loved this video once again! Ok here are my thoughts: Personalisation: I tried it and to be honest wont be using it again unless they change how it works. The fact that it constantly changes based on what I’ve liked as well as the ranking pairs which I think I ended up skipping around 90% of the pairs so I could sway the results - it took a LONG time. With including what you’ve liked on the explore page, just because I’ve liked something doesn’t mean I want to generate art like that. Additionally, I have several (relatively) consistent styles which are all different to each other that I try to achieve each month. When I can do that, I’ll happily jump over. Batch 8: I mean yay but also I wish it would go back to generating at like -q 0.5 and then picking something and running -q 5 to bump in more details. Image Editor: My gosh this is one of the best things to hit MidJourney!! Can’t wait to upload our own images. I do wish I could have the option of the rectangular select. I do use a tablet and while I love the brush off and brush on, drawing straight lines isn’t my forte lol Speaking of PNGs, I wish we could have transparent backgrounds. As a photographer/graphic artist, I’m excited about ai. Like you’ve mentioned you can either sit back begrudge the change or you can work with it to explore your creativity further. I think as creators we have a moral responsibility (as all creators ALWAYS have) in terms of copying etc. Even before AI you go onto Etsy and see how many people are ripping off IP from the likes of Disney etc - and funnily enough some of these people are the same people who are the first to complain about AI infringing on their art. They also probably fail to realise that while they don't mind ignoring the IP of big companies they are actually hurting the small businesses who abide by the rules. Explore page - I love following people’s personal explore page. I wish they would bring in an option to follow. I use my likes and a Notion board to keep track. Thank you for your time.
So many great insights here, thanks for taking the time to share the specifics. Super helpful to hear others opinions and experiences! And of course, thanks for tuning in. -DB
Hi guys, first of all I have to emphasize that for me this is the most valuable content regarding Midjourney. I have been working in this business for over 30 years and for the last two and a half years I have been intensively involved with the topic of AI and I just want to emphasize that I know what I am talking about with this statement ;-) I have a question for a current project that I am creating for a client. I am doing a series of famous people from the past who are shown in a certain situation. The focus here is on the similarity of the person. My current experience is as follows: some people, if they are very famous, are sometimes portrayed well, but this is the exception rather than the rule. Most of the time I have to fall back on a character reference with a portrait of that person. Now we come to the question: I am forced to work with version 6 because I get horrible results with version 6.1. The faces are far from the reference and look extremely artificial. So the question is, is there a way to achieve good results with the current version and if that is not possible, are there any tips to get as close as possible to a template with whichever version I can define via character reference or similar?
Happy to be here listening to you both. Loving it. One thing bugging me on the website ;- seems odd that we can’t add different weights to a bunch of your own images or say Pinterest images for image ref and style ref images using :: it’s frustrating not having the control of a few image weights or am I missing something.
Do you really think that personlization comes from the ranking thing? I honestly think It’s almost imposible to be so accurate. I think MJ uses our generations, otberwise there are thing that have no sense.
It’s a good thought. It could be true but as far as what they’ve continuously told us, it’s based on the rankings. I’m with you though-could totally see that. -DB
@@MidjourneyFastHours I know… but last week I generated over 500 images, a cristal ball with some things inside. Yesterday I ranked images and when I tested my new personalization code ( I use … - - p) one generation has a crystal ball. It can be a coincidence, of course. The only thing I know is that I love my personalization codes. Thanks for all guys! I feel less lonely.
I thought the ranking tasks were to select the most accurate image not your favourite image. Otherwise if I on;y create architectural stuff then I rank a group of cartoons how will that work. I would like to think -p is based on the explorer images you like cos there I’m very selective in liking interiors and architectural stuff. We need more clarity
Have questions for future episodes? Drop them below!
as usual so fun and so much good intel and ideas all the time!! Love you guys! playing with chemosynthetic for sure and giving me such good monsters for halloween -haha!
Thank you Drew & Rory! Really appreciate you guys taking the time to dig into my question! Yes, you got it. I want to take an already created piece of clothing and switch it onto a MJ styled model in Lifestyle & Runway imagery. I tested this extensively. MJ won't do it, but other programs will. #juxtapositioning #juxtaposin'
Thanks for the great question!
jux-ta-po….ah f it
Oh my gosh! What a start to your presentation. Sounds like you rocked it though! Loved this video once again! Ok here are my thoughts:
Personalisation: I tried it and to be honest wont be using it again unless they change how it works. The fact that it constantly changes based on what I’ve liked as well as the ranking pairs which I think I ended up skipping around 90% of the pairs so I could sway the results - it took a LONG time. With including what you’ve liked on the explore page, just because I’ve liked something doesn’t mean I want to generate art like that. Additionally, I have several (relatively) consistent styles which are all different to each other that I try to achieve each month. When I can do that, I’ll happily jump over.
Batch 8: I mean yay but also I wish it would go back to generating at like -q 0.5 and then picking something and running -q 5 to bump in more details.
Image Editor: My gosh this is one of the best things to hit MidJourney!! Can’t wait to upload our own images. I do wish I could have the option of the rectangular select. I do use a tablet and while I love the brush off and brush on, drawing straight lines isn’t my forte lol
Speaking of PNGs, I wish we could have transparent backgrounds.
As a photographer/graphic artist, I’m excited about ai. Like you’ve mentioned you can either sit back begrudge the change or you can work with it to explore your creativity further. I think as creators we have a moral responsibility (as all creators ALWAYS have) in terms of copying etc. Even before AI you go onto Etsy and see how many people are ripping off IP from the likes of Disney etc - and funnily enough some of these people are the same people who are the first to complain about AI infringing on their art. They also probably fail to realise that while they don't mind ignoring the IP of big companies they are actually hurting the small businesses who abide by the rules.
Explore page - I love following people’s personal explore page. I wish they would bring in an option to follow. I use my likes and a Notion board to keep track.
Thank you for your time.
So many great insights here, thanks for taking the time to share the specifics. Super helpful to hear others opinions and experiences!
And of course, thanks for tuning in.
-DB
Hi guys, first of all I have to emphasize that for me this is the most valuable content regarding Midjourney. I have been working in this business for over 30 years and for the last two and a half years I have been intensively involved with the topic of AI and I just want to emphasize that I know what I am talking about with this statement ;-)
I have a question for a current project that I am creating for a client. I am doing a series of famous people from the past who are shown in a certain situation. The focus here is on the similarity of the person. My current experience is as follows: some people, if they are very famous, are sometimes portrayed well, but this is the exception rather than the rule. Most of the time I have to fall back on a character reference with a portrait of that person. Now we come to the question: I am forced to work with version 6 because I get horrible results with version 6.1. The faces are far from the reference and look extremely artificial. So the question is, is there a way to achieve good results with the current version and if that is not possible, are there any tips to get as close as possible to a template with whichever version I can define via character reference or similar?
Happy to be here listening to you both. Loving it. One thing bugging me on the website ;- seems odd that we can’t add different weights to a bunch of your own images or say Pinterest images for image ref and style ref images using :: it’s frustrating not having the control of a few image weights or am I missing something.
Are you referring to when you're adding multiple --srefs and/or image refs?
@@MidjourneyFastHoursyes
Flux have said from day 1 the next release is their video module fyi!
Good to know
-DB
@@MidjourneyFastHours (see what's up next tab on their website for details :) )
Keep up the good work homies
@jzwadlo thanks brother ✌🏼
sref, cref and iref - The Trinity.
There we go!
🤘💪
nice
Do you really think that personlization comes from the ranking thing? I honestly think
It’s almost imposible to be so accurate. I think MJ uses our generations, otberwise there are thing that have no sense.
It’s a good thought. It could be true but as far as what they’ve continuously told us, it’s based on the rankings.
I’m with you though-could totally see that.
-DB
@@MidjourneyFastHours I know… but last week I generated over 500 images, a cristal ball with some things inside. Yesterday I ranked images and when I tested my new personalization code ( I use … - - p) one generation has a crystal ball. It can be a coincidence, of course. The only thing I know is that I love my personalization codes. Thanks for all guys! I feel less lonely.
@@ElenaFerrer_MeisiChannel interesting. Something to keep an eye on. Not out of the realm of possibilities.
I thought the ranking tasks were to select the most accurate image not your favourite image. Otherwise if I on;y create architectural stuff then I rank a group of cartoons how will that work. I would like to think -p is based on the explorer images you like cos there I’m very selective in liking interiors and architectural stuff. We need more clarity
35:34 multi prompting is wiiiiiiiildly undershared